Picture the intake you will do a hundred times in your first year of practice. A woman in her forties, nonseasonal major depression, moderate severity. She has been on an SSRI for ten weeks and is maybe a third better. You are already running the next step in your head: raise the dose, switch, augment with bupropion, maybe an atypical. Every one of those moves adds a side effect, a drug interaction, and a conversation about why the last thing didn’t work.
Here is the treatment almost nobody puts on that list. It has more than 30 randomized trials in depression. It sits in the American Psychiatric Association’s practice guidelines and the Veterans Health Administration’s. In the one head-to-head trial that matters most, it beat fluoxetine outright. It has no pharmacokinetic interactions, no weight gain, no sexual side effects, and you can’t overdose on it.
It is a 10,000-lux light box, and in the September 2026 issue of The Carlat Psychiatry Report, Dr. Norman Rosenthal, the Georgetown psychiatrist who published the first clinical description of seasonal affective disorder back in 1984, made the case that psychiatry has boxed this treatment into one diagnosis and one season for four decades while the evidence quietly outgrew both. If you are studying to become a PMHNP, this is one of those topics where a little knowledge separates you from most of the prescribers you will work alongside.
The trial that should have changed the conversation
Start with the study Rosenthal calls a landmark. In 2016, Raymond Lam and colleagues at the University of British Columbia published an eight-week randomized trial in JAMA Psychiatry that did something unusual: it tested bright light therapy in patients with nonseasonal major depression of at least moderate severity, and it tested it against an actual antidepressant. There were four arms of roughly 30 patients each: light alone, fluoxetine 20 mg alone, the two combined, and a double placebo of a sham ion generator plus a sugar pill.
The remission numbers are worth memorizing, because they are the kind of thing that reorganizes your thinking. Placebo: 30 percent. Fluoxetine alone: 19 percent. Light alone: 44 percent. Combination: 59 percent. Fluoxetine did not separate from placebo. Light did. And the combination produced a number needed to treat of about 3.5 for remission, which, if you have spent any time with antidepressant trial data, you know is an unusually good number.
One trial is one trial. That is why the paper Rosenthal points to next matters more. In a systematic review and meta-analysis published in JAMA Psychiatry in early 2025, Menegaz de Almeida and colleagues pooled 11 randomized trials with 858 patients, all with nonseasonal depressive disorders, comparing bright light therapy (alone or added to an antidepressant) against antidepressant alone, dim red light, or a sham device. Remission was 41 percent with bright light versus 23 percent without it. Response was 60 percent versus 39 percent. The American Psychiatric Association highlighted those figures on its own website as evidence that the treatment’s usefulness extends well past winter.
Rosenthal put the effect sizes in perspective for Carlat’s readers: large for seasonal affective disorder, around 0.84, and moderate for nonseasonal depression, around 0.53. That second number is roughly what a typical antidepressant delivers against placebo. That should not be an afterthought.

Why it works when there is no winter to blame
Here is where the topic stops being a trivia point and starts being physiology you will be tested on. The obvious question is: if light therapy treats a deficiency of daylight, why would it help someone who is depressed in July?
Rosenthal’s answer reframes what light is doing. Light, he told Carlat, is fundamentally an activating signal. Humans evolved to be active by day and quiescent at night, and depression can be understood partly as the physiology of night leaking into the day: low energy, psychomotor slowing, the machinery of sleep running while the person is supposed to be awake. Most of us now live indoors, where even a bright office delivers a few hundred lux; morning daylight delivers 10,000 to 20,000. Some people are simply more vulnerable to losing that signal than others, in any season.
The mechanism runs through the eyes, not the skin. Intrinsically photosensitive retinal ganglion cells containing melanopsin respond most strongly to blue-green light around 480 to 500 nanometers and project directly to the suprachiasmatic nucleus, the master circadian clock. From there the signal shapes melatonin timing, cortisol rhythm, and, through the raphe and locus coeruleus, serotonergic and noradrenergic tone. That is why morning light phase-advances the clock, improves alertness within hours, and why the Carlat table lists sleep-phase delay alongside SAD and nonseasonal depression as a condition with confirmed evidence.
A board question will not ask you to recite the retinohypothalamic tract, but it will test whether you understand that light therapy is a circadian intervention delivered through the eyes, that timing is part of the dose, and that it belongs alongside chronotherapy rather than being filed under “alternative medicine.”
The prescription, written out
The reason light therapy gets underused is not that clinicians doubt it. It is that most were never taught to prescribe it with the precision they bring to a starting dose of sertraline. Rosenthal’s protocol in the Carlat interview fixes that.
The standard is a 10,000-lux box used in the morning, starting at 20 to 30 minutes per session. If there is no benefit after two weeks, extend toward 60 minutes, or add a late-afternoon session for patients who crash around 4 to 6 pm. Never after 8 pm, because evening light delays the clock and wrecks sleep. The box needs a UV filter and a screen around 12 by 17 inches, positioned a foot or two in front of the face, at or slightly above eye level, angled toward the patient. Closer than 12 inches strains the eyes; staring directly at the panel is not necessary. The light has to reach the retina, so a box beside the monitor while the patient reads email is fine, and a box on the shelf behind them is not.
Rosenthal’s practical tip is the one your patients will actually use: attach the session to something they already do every morning. Coffee, email, breakfast. His own light box sits above his computer monitor. Adherence is the whole game with this treatment, and adherence is a design problem, not a willpower problem.
Timing of response differs by population. In seasonal affective disorder, patients often improve within days, sometimes hours; Rosenthal notes that a response in the first hour predicts response at one week. In nonseasonal depression, the benefit builds over two to three weeks, closer to the antidepressant timeline. Set expectations accordingly. And for patients whose winter depression is predictable, start before they need it, at the first prodromal signs: less energy, lower libido, a few extra pounds, more appetite for carbohydrates. Not the first PHQ-9 above 10.

What can go wrong, and who should not use it
A treatment with almost no side effects is still a treatment, and your boards will look for whether you know the exceptions.
The one that matters most is bipolar disorder. Light is activating, and in a patient with a bipolar diathesis, activation can tip into hypomania or mania. Light therapy does have positive evidence in bipolar depression, but the protocol is different. The Carlat summary is specific: titrate slowly, starting at 15 minutes a day and increasing by 15 minutes per week, and dose at midday, between noon and 2 pm, rather than first thing in the morning. Midday light delivers the antidepressant effect with less of the phase-shifting punch that seems to drive switching. If a patient has any history of mania, that is the version you use, with a mood stabilizer on board.
The second category is ocular. Patients with glaucoma, cataracts, or retinopathy, including diabetic retinopathy, should use light therapy only under the supervision of an ophthalmologist. This is not a contraindication, but it is a referral you need to make before recommending the box.
The third is photosensitivity. Systemic lupus and photosensitizing medications call for caution. The Carlat table specifically flags lamotrigine, worth remembering because lamotrigine is a first-line agent in the very bipolar depression population where you might be considering light.
Ordinary adverse effects are headache, eye strain, and nausea, usually managed by shortening the session or moving the box farther away. Rosenthal also names the failure modes he sees when light therapy “doesn’t work”: the box is off to the side, it is a small travel-sized unit that only delivers 10,000 lux at three inches, it is being used at night, or the dose is too low. Fix positioning and dose before you conclude the treatment failed.
What the evidence does and doesn’t support
Part of being a good clinician is knowing where the data runs out, and Rosenthal is candid about it. The confirmed list, meaning six or more randomized trials, is seasonal affective disorder, nonseasonal depression, and delayed sleep phase. The “possible” list, with one to five trials, includes bipolar depression, depressive symptoms in anorexia, agitation and sleep in dementia, PTSD, and negative symptoms in schizophrenia. There are also around seven randomized trials in children and adolescents; Rosenthal’s own NIMH study with Susan Swedo used 2,500 lux for children under nine and the full 10,000 for teens.
Dawn simulators, which ramp bedroom light to 100 to 300 lux over 30 to 120 minutes before wake time, have treated SAD on their own in nine small trials. Their particular value is in sleep inertia, the heavy grogginess that keeps SAD patients from getting to the light box at all. For an adolescent who cannot get up for school, a dawn simulator is a nonpharmacologic first move.
Now the things he does not recommend, because patients will ask. Light visors: his own NIH lab studied them extensively and could never show a dose-response relationship. Red light for mood: the data behind the marketing is thin. An outdoor walk instead: fine in summer, when morning light reaches 10,000 lux or more, but near the winter solstice an hour outside will not deliver the dose, and it makes treatment dependent on the weather. Modern LED boxes appear as effective as fluorescent ones and don’t degrade over years. The specifications to give a patient are simple: 10,000 lux, UV-filtered, screen at least 12 by 17 inches, white light.
Why this matters for you specifically
PMHNPs are trained as prescribers, and the reflex your program builds is a pharmacologic one. That reflex is valuable. It also has a blind spot, and light therapy sits squarely in it. The Carlat interview’s closing line is blunt: if you are not offering it, your patients are missing out.
In the medically complex patient on six drugs where every augmentation carries an interaction, light is a clean add-on. In the pregnant patient where every pill is a risk-benefit conversation, light has trial data in perinatal depression. In the adolescent whose parents are wary of medication, it is something concrete to offer at the first visit.
For your boards, the testable core fits on an index card. Ten thousand lux, morning, 20 to 30 minutes, titrate to 60. Effect size large in SAD, moderate in nonseasonal depression, comparable to antidepressants. Works through the retina and the suprachiasmatic nucleus, not the skin. First-line for SAD even when severe, with bupropion the only FDA-approved medication for prevention of seasonal episodes. Mania risk in bipolar patients: start low, go slow, dose at midday. Eye disease needs ophthalmology. Photosensitivity: watch lupus and lamotrigine. Not after 8 pm. Any one of those is a stem waiting to happen.
The larger lesson is about how you will think for the next thirty years. The prescribers who get the best outcomes are not the ones with the longest formulary. They are the ones who understand mechanism well enough to recognize a good tool when it doesn’t come in a bottle. If you want to build that kind of clinical reasoning into how you prepare for the ANCC or AANP psychiatric-mental health boards, come see how we teach it at PMHNPReview.org. Memorizing the lux number takes a minute. Learning to look at a stalled SSRI and think “circadian” takes a little longer, and it starts today.
